New Data & Democracy Research Initiative Launched at University of Chicago
Democracy faces increasingly worrisome threats in today’s digital and online world: The Internet and social media amplify the spread of misinformation, conspiracy theories, and extremist political rhetoric, intensifying political polarization and bias. Utilization and manipulation of online spaces has profound impacts on offline political activity, from voting to civil unrest. Twenty-first century civic and political engagement are defined by online platforms’ data security, information flow, and moderation policies. Understanding and addressing these challenges and their effect on governments and society requires new inquiry and expertise at the intersection of democracy, public policy, technology, and data science.
The Data & Democracy Initiative, a new research initiative at the University of Chicago, will ignite interdisciplinary research on the digital challenges facing democracies around the world. A joint collaboration of the UChicago Data Science Institute (DSI) and the Center for Effective Government (CEG) at the Harris School of Public Policy, the initiative will seed new collaborations across the social and computational sciences, producing research-driven reports, tools, and software that bolster the work of policymakers, social impact organizations, and communities working to strengthen democracy.
“The most pressing questions facing political scientists today are intertwined with an ever-evolving online ecosystem and near-infinite amounts of data. Understanding and strengthening democracy in service of a more effective government requires bridging divides between the online and offline worlds—and the methods used to study them,” said William Howell, CEG Director and Sydney Stein Professor of American Politics at UChicago. “The Data & Democracy Initiative will do just that, leveraging world-class faculty with expertise across disciplines alongside skilled practitioners and leaders to bring into sharp relief the unique challenges facing democracy in the digital age, and their potential solutions.”
The initiative’s first activity is an inaugural round of research funding, seeding four new interdisciplinary collaborations addressing questions on political participation and freedom of expression in the digital age. These projects bring together experts in policy, politics, law, computer science and data science to create new approaches for analyzing social media content moderation, neurological signals of political bias, the relationship between political polarization and dysfunction, and the effect of messaging on online information consumption and behavior. Among the grantees are UChicago CS faculty Marshini Chetty, Nick Feamster, and Chenhao Tan.
The initiative will ultimately build a community of scholars and students at the intersection of democracy and data science, establishing new opportunities and expanding upon existing activities at both CEG and DSI. The initiative is currently seeking a postdoctoral scholar to contribute to interdisciplinary, collaborative work on cutting-edge research projects, with application review beginning March 21. (Interested candidates are encouraged to apply here.) Additionally, several recipients of the initiative’s initial round of research funding will serve as mentors in DSI’s Summer Lab—an immersive 10-week paid summer research program for high school, undergraduate, and masters students—working with students specifically on Data & Democracy projects.
In addition to stimulating original research, the Data & Democracy Initiative will also hold public events and build new partnerships with academic, non-profit, and private organizations and practitioners focused on fighting disinformation and other digital threats to democracy. These outreach efforts will help disseminate research findings to the entities engaged in strengthening democracy, providing them with conclusions and interventions supported by scientific evidence. As part of these efforts, CEG announced last month that two of its four 2022 Senior Practitioner Fellows—Tiana Epps-Johnson and Mindy Finn—would focus exclusively on the work of the Data & Democracy Initiative during their yearlong fellowship with the Center.
“The Data & Democracy initiative builds on the research vision of the UChicago Data Science Institute to tackle interdisciplinary, data-driven problems that have real-world, societal impact,” said Nick Feamster, Neubauer Professor of Computer Science and Faculty Director of Research at DSI. “The threats to free and open communication online have profound effects on our political systems, and on the flip side, our governments play a huge role in our ability to find and exchange information. Data-driven methods at the intersection of computer science, law, and policy can play a valuable role in understanding, and ultimately developing solutions for, the challenges we face in this realm.”
Those interested in learning more about the Data & Democracy Initiative or getting involved in its work can contact the Data Science Institute at data-science@uchicago.edu or the Center for Effective Government at effectivegov@uchicago.edu
For more on the inaugural Data & Democracy Initiative grants, see below:
Partisan Polarization and Political Dysfunction
William Howell (Political Science), Anthony Fowler (Harris School of Public Policy), Andrew Eggers (Political Science), Chenhao Tan (Computer Science)
Political polarization and dysfunction often occur together, but questions remain about the causal relationship between these phenomena. Does polarization produce government dysfunction, or do impediments to lawmaking propagate political extremism? This project will connect political scientists and computer scientists to evaluate and refine the understanding of this relationship through empirical investigations. Topics of interest include instances where the capacity of the state or a state office experience a sudden, exogenous shift in the ability to meet a policy challenge, and the downstream electoral consequences of changes in the constitutional design of an entire system of government. The research will also use natural language processing techniques to build new measures of candidate policy positions and speech. Results will help democratic reformers determine the most effective path to protect against political extremism and impediments to government action.
Understanding and Auditing Content Moderation Policies
Marshini Chetty, Chenhao Tan, Nick Feamster (Computer Science), Genevieve Lakier (Law)
Internet and social media platforms face growing challenges in moderating the content posted by users, deploying a range of policies to find, flag, and remove offensive or inappropriate material. This study will combine computer science and law approaches to conduct an empirical study of the moderation policies and practices of platforms such as Youtube, Reddit, Facebook, and Etsy, creating the first public corpus of these policies and auditing how these systems are applied in real situations. The researchers will also conduct surveys and interviews to assess how users of these platforms perceive these policies and their expectations for how content rules will be enforced.
Understanding Divergent Interpretations of Political Information
Yuan Chang Leong (Psychology), Daniel Grzenda (Data Science Institute)
The current atmosphere of politically polarized media has strengthened individuals’ political bias in how they interpret news and other information. Using a combination of methods from neuroscience, linguistics, and computer science, this collaboration will refine existing approaches for measuring political bias. fMRI studies will measure the neural responses of conservatives and liberals while watching political debates, while natural language processing methods will build predictive models of the types of content that increase these divergent brain responses. The results will help create interventions that combat messaging designed to further inflame polarization, facilitating healthier political discourse.
Facebook Messenger Chatbot for Surveys and Intervention Testing
Molly Offer-Westort (Political Science), Nick Feamster (Computer Science)
For many Internet users, Facebook is the portal to the Internet. Understanding how people consume and engage with information on Facebook is central to understanding how information and misinformation spreads on the Internet and shapes perspectives. Chatbots offer a promising new research tool for engaging with users within the context of their regular social media engagement. This project will develop a new Facebook Messenger chatbot that can be used for a variety of research uses: canvassing on a broad range of topics, measuring information consumption, or studying the effect of messaging interventions on political participation and behavior. The researchers will also use adaptive, data-driven algorithms that learn from exchanges with users and determine optimally persuasive messaging around political events.
